v1.2.0

Add Prometheus metrics export

Implements full Prometheus observability integration with /metrics endpoint.

Features:
- 11 Prometheus metrics covering dumps, storage, restarts, and operations
- Background metrics collector (updates every 30s from database)
- Real-time metrics recording on dump upload, prune, and cleanup
- Proper label support (namespace, pod, deployment) for filtering

Metrics added:
- oom_watcher_dumps_total{namespace,deployment} - Total dumps captured
- oom_watcher_dumps_by_namespace{namespace} - Current dumps per namespace
- oom_watcher_memory_percent{namespace,pod} - Memory % at dump time
- oom_watcher_avg_memory_percent - Average memory across all dumps
- oom_watcher_time_since_last_dump_seconds{namespace,pod} - Time since last dump
- oom_watcher_storage_bytes - Total storage used
- oom_watcher_pod_restarts_total{namespace,pod} - Pod restarts observed
- oom_watcher_prune_operations_total - Prune operations executed
- oom_watcher_pruned_dumps_total - Dumps deleted by pruning
- oom_watcher_pruned_bytes_total - Bytes freed by pruning
- oom_watcher_cleanup_operations_total - Cleanup operations executed

Helm chart updates:
- Add Prometheus scrape annotations to Service and Deployment pods
- Add optional ServiceMonitor support for Prometheus Operator
- Configurable metrics collection (enabled by default)
- Fix: Use targetPort (8080) instead of service port (80) for scraping
- Bump chart version to 0.2.5

Documentation:
- Add comprehensive Prometheus section to README
- Include scrape config examples and alert rules
- Update roadmap marking feature complete
- Add Grafana dashboard suggestions

Other:
- Remove automatic 'latest' tag from build-and-push.sh script

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>